Two scripts vernacular

DAB is great. Had a dictionary app in less than 2 hrs. So thanks!
The language I work with has two scripts: Cyrillic and Latin. This is why the LIFT has two language version for lots of items (e.g. paradigm entries, examples). Most go well. There is a problem, however, for the lexical unit. E.g:

<lexical-unit>
<form lang="ce"><text>хьаьрмак</text></form>
<form lang="ce-Latn"><text>hwaermak</text></form>
</lexical-unit>

The ‘details view’ of this goes well, no problem there.
The problem I have is with the top-level interface where the user can choose the main language for searching and viewing. The Cyrillic nicely shows the correct Cyrillic entries. The Latin one fails. It doesn’t show the ‘lexical-unit’ entries. What it shows instead is the few (four) entries that have a definition field in [ce] and [ce-Latn].

The app-builder “Languages” tab correctly lists a number of languages as type “Gloss”, then [ce] has type “Vernacular”, and [ce-Latn] has two types “Vernacular, Gloss”.

I can understand that for the other languages the definition fields and gloss fields (if they are used) should be used as a basis for building indexes and browsing through the dictionary.

But is it possible to have a dictionary view based on [ce-Latn] too?

Thanks in advance!

You can with LIFT add a reversal that will generate an index for the ce-Latin.

If you know RegEx it may be doable that way but I would do it in XSLT. I have some scripts that can do a bit similar to that.

You’re right, that’s the solution.
The LIFT representation of the dictionary should have <reverse> entries with the language set to ce-Latn in my case.

I was able to do that with a combination of Flex, reverse entries and notepad++.
Thanks for pointing into the right direction!